Conventional online poker tracking software provides player behavior statistics based on averaging all the hands on record for a player, or for the current session. These statistics are limited in that they do not provide summaries of recently played hands within a session, and precise histories of single hands included in those recently played hands. Further, known poker tracking software requires an extra, non-automated, configuration step in which the burden is on the user to specify values that determine classifications of player behavior. Still further, conventional notifications to a user of specific play information relative to an opponent is limited in that the information is based on a percentage of time that the opponent makes that specific play. Because of the limitations and deficiencies described above, there exists a need for an improved technique for monitoring an online card game.